IRON MOUNTAIN DATA CENTERS
STATS :
• To train a ChatGPT 3.5 model requires 300-500MW of power
• With generative AI in the mix , data centre CAGR could rise as high as 25 %, tripling capacity in five years and increasing it up to tenfold in a decade
• By 2030 , annual e-waste production is on track to reach a staggering 75 million metric tonnes
• The IMDC Asset Lifecycle Management ( ALM ) division , which now covers 32 countries , sanitises more than three million drives a year and has generated in excess of US $ 1bn for clients via remarketing and recycling
• Click here to read the IMDC Infrastructure Service Sheet on the Top 10 considerations when planning AI infrastructure
the infrastructure industry that supports it . It promises to deliver immense economic value over the coming decade , but will also consume immense amounts of power . Many generative AI applications can be hosted in a specialised shared facility . Different models have different infrastructure requirements , but all share the need for high-density power , advanced cooling and modular design .
The scale of the power challenge does not mean it cannot be overcome . In an era in which Big Tech has displaced many oil giants in the list of the world ’ s largest companies , innovation has accelerated and serious fact-based commitments have been made by industry leaders to tackle the climate crisis . In fact AI may be a key to solving the problem , not just for our own industry but for other sectors .
Data centre customers interested in developing generative AI applications should plan and invest early to keep ahead of the steep upward curve in power and space uptake . They should also pay close attention to new infrastructure design impacts ,
170 December 2023